A novel approach to potato processing using intelligent data analytics and advanced control

在食品行业食品的生产厂家，有着硬性的操作标准和加工环节 必须遵守的方法论。在许多情况下，这些操作和处理方法需要 它们被很好地控制在严格的公差范围内。加工的生土豆种类繁多， 产品只是一个例子。McCain Foods自1957年以来一直是加拿大NB Florenceville的一家公司， 致力于不断改进和应用新技术，以生产近 世界上三分之一的法式炸土豆该项目的重点是进行研究，以便 提高马铃薯去皮过程，这是马铃薯薯条生产整体效率的关键。 目前，剥离过程涉及使用高流量的蒸汽进入旋转容器，所述旋转容器包含： 原材料在规定的时间内。蒸的最佳时间在这一点上还不知道， 船的旋转速度。其他参数影响去皮过程，如马铃薯尺寸 和类型。由于蒸制过程是生土豆去皮的关键， 该初始过程用于最小化能量消耗和最大化剥离效率。过度蒸煮或 在汽蒸条件下，可能导致剥皮不良，每年造成数百万美元的收入损失。的 研究将涉及几个研究人员进行顺序和并发的任务，包括 下面：（a）根据第一原理制定整个马铃薯剥皮机构的分析模型， (b)进行各种实验以制定非线性时间序列模型，（c）整合高分辨率 热成像摄像机，其在汽蒸期间和之后提供原料的热特征，（c） 使用热数据来实时增强这些时间序列模型，以及（d）开发智能的 控制器使用这些信息（热信息和许多其他信息）来对多个致动器做出最佳动作 （阀门、加热器、电机等）。麦凯恩食品公司将提供5万美元现金和5.44万美元实物 支持这个项目。该项目的时间轴为2年。

In the manufacturer of food in the food industry, there are rigid operational standards and processing methodologies that have to be adhered to. In many instances, these operations and processing methods demand that they are well controlled within tight tolerances. The processing of the raw potatoes to a wide variety of products is just one example. McCain Foods in Florenceville, NB, Canada since 1957 as a company has been committed itself to be continuously improving and applying new technologies to its production of nearly one-third of the French fried potatoes in the world. This project focuses on conducting research in order to enhance the potato peeling process, which is key to overall efficiency of the production of potato fries. Currently, the peeling process involves using steam at high flows entering a rotating vessel that contains the raw material for a defined period of time. The optimal time for steaming is not known at this point as well as the vessel's rotating speed. Other parameters influence the peeling process such as the potato dimensionality and type. As the steaming process is key to the peeling of the raw potato, the company is keen on improving this initial process for minimizing energy consumption and maximizing peeling efficiency. Over steaming or under steaming can result in poor peeling, translating to millions of dollars annually in lost revenue. The research would involve several researchers conducting sequential and concurrent tasks that include the following: (a) formulating from first principles an analytical model of the overall potato peeling mechanism, (b) conducting various experiments to formulate nonlinear time series models, (c) integrating high resolution thermal imaging cameras that provide thermal signatures of the raw material during steaming and after, (c) using the thermal data to enhance these time series models in realtime, and (d) developing an intelligent controller that uses this information (thermal and many others) to make optimal actions to the several actuators (valves, heaters, motors etc.). McCain Foods will be providing $50,000 in cash as well as $54,400 inkind support on this project. The project has a timeline of 2 years.